Abstract
As Unmanned Aerial Vehicles (UAVs) become integral to modern airspace, providing reliable wireless connectivity via terrestrial cellular infrastructure presents unique challenges. This paper investigates two service provisioning schemes: the Inclusive-Service (IS-BS) scheme, where base stations serve both ground and aerial users, and the Exclusive-Service (ES-BS) scheme, which uses dedicated base stations. By optimizing antenna tilt angles, the study identifies critical network density thresholds that determine which scheme minimizes network outage probability.
The Challenge of 3D Connectivity
Traditional cellular networks are optimized for a 2D world. Base station (BS) antennas are typically tilted downward to maximize signal for ground users (GUs) and minimize inter-cell interference. However, high-flying UAVs, or aerial users (AUs), often reside in the antenna's side lobes or nulls, leading to weak signals. Furthermore, the clear line-of-sight (LoS) between UAVs and distant base stations can lead to significant interference. This research addresses whether existing infrastructure should be shared or if dedicated resources are necessary for aerial connectivity.
Service Provisioning Schemes
The researchers compare two fundamental architectures to address these challenges:
- Inclusive-Service (IS-BS): Every BS simultaneously serves both GUs and AUs. This maximizes the number of potential connection points for all users.
- Exclusive-Service (ES-BS): BSs are partitioned into two groups, one dedicated to GUs and the other to AUs, allowing for specialized antenna tilts tailored to each group's elevation.
Methodology
The study employs stochastic geometry (specifically a Homogeneous Poisson Point Process) to model the random distribution of base stations and users in 3D space. It incorporates realistic 3GPP antenna patterns and distance-dependent LoS probabilities to capture the complex propagation environment UAVs encounter. The core metric analyzed is the network outage probability, derived using Laplace transforms to handle aggregate interference from randomly located base stations.
Optimizing Antenna Tilt and Network Density
Adjusting the antenna tilt angle involves a "conflicting impact." Decreasing the downward tilt to better serve UAVs expands the main lobe's service area (a positive effect) but increases the link distance to the serving BS (a negative effect). This creates a U-shaped optimization problem for finding the ideal tilt angle.
A key finding is the existence of a "Critical Density" ($\lambda_B^c$). In sparse, noise-limited networks, the IS-BS scheme is superior because it offers more connection candidates. Conversely, in dense, interference-limited environments, the ES-BS scheme performs better by using specialized tilts to isolate user groups from cross-interference.
Key Results
- Optimal Tilt Existence: Numerical verification confirms that an optimal tilt angle exists for both user types across various environments.
- Density Impact: In high-density scenarios (e.g., $\lambda_B \geq 2 \times 10^{-5} \text{ BSs/m}^2$), the ES-BS scheme significantly reduces outage probability. In low-density scenarios, the IS-BS scheme can reduce outage by 10-15% due to macro-diversity.
- Height Sensitivity: As UAV altitude increases, the optimal tilt for shared systems must shift upward to compensate for path loss. However, as the proportion of ground users increases, the optimal tilt in a shared system must shift further downward to protect the majority of users on the ground.
Conclusion
The choice between shared and dedicated base station infrastructure for UAVs depends heavily on network density and interference levels. While dedicated stations offer better performance in crowded urban environments by mitigating interference, a shared approach remains more efficient for sparse deployments. Future work will need to address uplink interference and the impact of UAV mobility on dynamic beamforming.
Summary updated